Instapundit (second update) refers once more to Paul Krugman's 'rhinoceri'. That's as bad as 'octopi'. If you want to be pedantic and use the Greek plural, they're 'rhinocerotes' (five syllables) and 'octopodes' (four syllables). If you want to be really pedantic, you can even use the Greek pronunciations and call them 'REE-no-keh-ROW-tace' and 'ock-TOP-o-dace'. But who would want to do that? Certainly not I.

Better to stick with the standard English plurals 'rhinoceroses' and 'octopuses'. If these sound too ugly, we can pretend that the plural is the same as the singular and say 'a herd of rhinoceros' or 'a plateful of octopus'. That works with squid and haddock and moose and trout, why not rhinoceros(es) and octopus(es)? Dictionaries may object, but my ears do not.

Krugman's meaning is also obscure, as he must explain that it refers to Ionesco's play, which most of us haven't read. What is particularly confusing in a political context is that many Americans (especially on Free Republic) already use RINO to refer to a 'Republican In Name Only' like John McCain, or Jim Jeffords before he came out of the closet as a Democrat. Of course the same thing on the other side is a DINO. When Krugman calls Mickey Kaus a rhinoceros, he means what others would call a DINO, not a RINO.

Just to further confuse things, ancient Romans used rhinoceros as "a nickname for a man with a long nose" and metaphorically for a Frasier Crane or Gil Chesterton type of critic: in the epigrammatist Martial, nasum rhinocerotis habere, "to have the nose of a rhinoceros" means "to turn up the nose, to sneer at every thing". Does that make Krugman a rhinoceros when he writes about Republicans?

Update: (8:40 PM)

In the first comment, 'barbara' quote Webster's as approving 'octopi'. I think Webster's ought to be ashamed. I object to 'octopi' because it's neither Latin nor English, it's pseudo-Latin. It seems to me perfectly proper either to use the original foreign plural of a foreign word, or to use the English plural with 's' or 'es'. But if you're going to use the original foreign plural, you need to get it right. Thus, the plural of 'cherub' and 'seraph' could be either 'cherubim' and 'seraphim' (Hebrew) or 'cherubs' and 'seraphs' (English), whichever sounds better. But making them 'cheruboi' (Greek) and 'seraphen' (German) would be wrong.

In Latin, words of the same ending do not always have the same plural. The plurals of alumnus and genius (2nd declension) are alumni and genii, but the plural of genus (3rd declension neuter) is genera, the plural of Venus (3rd declension feminine) is Veneres (= 'statues of Venus'), and the plurals of census, hiatus, and apparatus (4th declension) are census, hiatus, and apparatus (with a long U in the last syllable instead of a short). I was going to correct James Lileks' inexcusable 'hiatii', approved by VodkaPundit, but Sightseeing in Plato's Cave beat me to it.

In some cases, either the Latin or the English plural sounds good: 'indexes' or 'indices', 'appendixes' or 'appendices', take your pick. (We can observe a general trend towards the English forms with X as fewer people take Latin.) In other cases the Latin plural sounds much better: no one would say 'genuses' for 'genera', 'axises' for 'axes', or 'basises' for 'bases', even though the last two are easily confused with the plurals of 'axe' and 'base'. That is no doubt why some foreign plurals survive. But adding any old Latin ending to any Latin word seems wrong, no matter what Webster's says. And I don't think it's just my personal whim that says 'octopi' is wrong: it's not Latin, it's not Greek, it's not English, what excuse does it have?

Similarly, English-speakers are under no obligation to sprinkle their speech with French phrases like 'je ne sais quoi' and 'soupçon'. But those who do will want to spell and pronounce the words correctly. (If I haven't, please let me know: I'm only quoting these words as part of my argument, and may well have gotten them wrong.)